Does gesture type matter? Does iconicity matter?

The aim of this section is to analyze whether the experimental evidence puts forward a specific advantage of adding different types of manual gestures to the learning of new words. Specifically, one could hypothesize that if the gesture puts forward an iconic resemblance with the referent it could be more beneficial for learning the new word labeling the ref-erent: “if a sign is more iconic than the spoken word, its form conveys information about a word’s meaning and may thus assist a child in map-ping new words to meanings” (Bird et al., 2000, p. 260).

Lüke and Ritterfeld (2014) found that both iconic and arbitrary gestures were equally beneficial to word learning in TD children aged 4;9 years.

This could appear as contrary to the results of Namy and Waxman (1998) who compared the ability of children from 18 to 26 months to learn either word or arbitrary gestural labels. They found that 18-month-olds learned word or gestural labels indifferently whereas 26-month-olds learned word labels more easily and needed extra training to learn gestural labels. Bird and colleagues (2000) also found that 21.8-month-olds were not able to expressively learn arbitrary gestural signs alone as labels. Marentette and Nicoladis (2011) found that children aged 40 to 60 months could learn

iconic gesture labels for objects but not arbitrary gesture labels. This dif-ference in findings could suggest that, even if children have difficulties learning only an arbitrary gestural label (without a word) for an object, the arbitrary gesture could however help them learn a corresponding and simultaneously presented word (at least receptively). Lüke and Ritterfeld (2014) hypothesize that “arbitrary gestures may have enhanced the interest of the child in the presented words in contrast to words intro-duced without gestures”. This finding is however contradictory to that of Bird and colleagues (2000) who found no difference between a word only and a word + arbitrary sign condition on expressive and receptive word learning (mean age = 21.8 months). This difference in findings could be due to the age of the participants: the ones in the Bird et al. (2000) study may have been too young to manage to learn arbitrary signs even though Namy and Waxman (1998) found that 18-month-olds can learn words and arbitrary gestural labels. This hypothesis is backed by the fact that Bird and colleagues (2000) did find a beneficial effect of arbitrary signs for receptive learning of new words in older children with T21 (mean age = 42.3 months). Note however that the participants with T21 were trained in using signs prior to the study whereas TD children were not.

Giezen, Baker and Escudero (2013) found no benefit in adding arbitrary signs to learn new words in 6;11-year-old children with CI. The scores were however close to 100 % and the absence of a gesture benefit could be due to a ceiling effect.

Another surprising observation, taking into account the results of Marentette and Nicoladis (2011), is that Lüke and Ritterfeld (2014) found no advantage of iconic gestures over arbitrary ones. This could be due to a ceiling effect for iconic gestures (as suggested by the authors themselves).

Vogt and Kauschke (2017a) ran an interesting follow-up experiment to their main study comparing iconic and attention-directing gestures. It compared the use of iconic gestures versus arbitrary ones. Even though sample size was small (18 TD children) and impeded reaching statistical significance (according to the authors themselves), the results suggest that both expressive and receptive performances were higher for the iconic than the arbitrary gesture condition.

Mollink, Hermans and Knoors (2008) found a positive effect of adding signs to words for receptive word learning in hearing impaired children.

They analyzed this effect as a function of sign iconicity. Even though the effect was the same for signs with strong iconicity than for those with weak iconicity one week after training, the results after five weeks show that learning was better for strongly iconic signs. The interesting thing is that at five weeks, the performance decreased compared to one week only for weakly iconic signs but remained the same for strongly iconic signs suggesting that iconicity helped longer memory retention. This finding is corroborated by that of van Berkel-van Hoof and colleagues (2016) who found a positive effect of adding iconic signs for receptive word learning in hearing impaired children. Note however that they did not find a bene-ficial effect of iconic gestures over none in 10;8-year-old TD children and children with SLI. Lüke and Ritterfeld (2014) did find a beneficial effect of iconic gestures over none in 4;7-year-old children with SLI.

Capone and McGregor (2005) and Capone Singleton (2012) com-pared the effect of using iconic gestures underlying shape to ones under-lying function. They found that shape gestures were more efficient in promoting expressive and receptive word learning than function gestures.

This suggests that type of iconicity could be as important as iconicity itself.

In both studies the shape gestures were static symbols and the function gestures were dynamic symbols. Even if performances on words trained with a function gesture were generally not better than those for words trained with no gesture, the authors put forward an interesting finding. In expressive learning testing, when the participants did not manage to pro-duce the word, the experimenter provided the gestural cue. In these cases, function gestures functioned as good as shape gestures to help the children produce the new words upon testing, suggesting that function gestures may have a beneficial effect even though they are not as effective as shape gestures. Capone Singleton (2012) speaks of a shape bias already put for-ward by other researchers (Kemler Nelson et al., 2000; Smith et al., 2002).

Mumford and Kita (2014) also compared different types of iconic gestures for learning new verbs receptively: ones underlying the manner of an ac-tion (dynamic) and the other its resulting end state (shape, static). Contrary to the above-mentioned studies, the authors found a beneficial effect over word only training solely for the dynamic manner gestures and not for the static end state gestures. This may be due to a difference in the ages of the participants: the participants were aged 41.48 months on average whereas

those in the Capone studies were aged around 28  months. Actually, it appears that the shape bias mentioned above would wear off with age (Imai, Gentner, & Uchida, 1994). Also note that in the Mumford & Kita (2014) study, the children learned verbs whereas they learned nouns in the Capone studies. Vogt and Kauscke (2017a) compared iconic gestures underlying path and/or manner to ones underlying shape. They found a larger advantage for path-manner than shape gestures for immediate learning of verbs but similar effects for both gesture types for immediate learning of nouns. On the other hand, after a two- to three-day delay from the end of training, they found a larger advantage for shape than path-manner gestures for nouns but similar effects for both gesture types for verbs. This result helps understand the differences between the Mumford

& Kita study and the Capone studies.

Vogt and Kauscke (2017a) found an advantage of iconic gestures over attention-directing gestures (raised forefinger in front of upper body) even at the end of the first training session. The authors conclude that: “it is the iconicity of the gestures (that is the resemblance to the referent), rather than the item-specific encoding of both auditory and visual information to a lexical form, that helps learning” (p. 22). O’Neill, Topolovec and Stern-Cavalcante (2002) analyzed the generalization of the use of newly learned adjectives to qualify other objects than those used during training with a similar distinctive quality referred to with the adjective. For example, during training, the children were presented with a ‘lumpy cat’ and taught the adjective ‘lumpy’. Upon testing they were presented with a ‘lumpy turtle’ and a ‘smooth turtle’ and asked to designate the ‘lumpy’ one. During the learning phase, some adjectives were learned with a descriptive gesture and others with a pointing gesture. In a first experiment, they found no difference between the two conditions except for the adjectives describing non-visual properties of objects (descriptive gesture advantage) and con-cluded that: “gesture may play a more important role in the learning of less visually detectable properties” (p. 255). In a second experiment, using lower frequency adjectives describing only non-visible properties, they found that descriptive gestures were more efficient in promoting learning than pointing gestures. An interesting finding is a higher frequency of men-tion of nontarget properties (resp. less expression of uncertainty during testing) by the participants in the pointing than in the descriptive gesture

condition, but only in experiment 1. Capone Singleton (2012) compared iconic gestures underlying shape to pointing. She found that shape gestures were more efficient in promoting expressive and receptive word learning than pointing. Booth, McGregor and Rohlfing (2008) however found that the use of pointing by the experimenter during training yielded better receptive learning than not using any gesture. It may therefore be the case that attention-directing gestures are also beneficial for word learning, but less than iconic gestures.

Overall, the studies reviewed here suggest that different types of manual gestures can have differential effects. Pointing gestures appear to be helpful for word learning compared to no gesture even though there is some evidence that gestures having an iconic resemblance with the ref-erent would be more effective. “the use of descriptive gestures during the teaching of novel adjective terms appears (…) to have helped children to isolate the particular property intended by the speaker in a manner not possible when point gestures were used instead” (O’Neill, Topolovec, &

Stern-Cavalcante, 2002).

Results of the studies comparing arbitrary and iconic signs do not all agree but do suggest in general that iconics are more beneficial. Several studies also put forward differences between iconic gestures with a bias of shape over function for nouns and path-manner over shape gestures for verbs. Capone and McGregor (2005) suggest that the role of gestures is to draw “attention to an important aspect of the word learning problem (shape, function or both), thereby reinforcing salient semantic content of the spoken language” (p. 1478). As suggested by Vogt and Kauschke (2017a), iconic gestures may facilitate the association with the lexical form. Type of iconic gesture could interact with type of word learned (e.g., nouns vs. verbs) and also the word itself. For example, one could hypothe-size that some words are better represented by static shape iconic gestures (e.g., hands shaped as a round to illustrate the word ‘ball’) and others by dynamic function gestures (e.g., fingers miming cutting to illustrate the word ‘scissors’).